Nvidia says the US government granted it a license to ship a small number of H200 chips to Chinese customers, but it's unclear if Beijing will allow any imports
Context & Ripple Effects
The license follows Washington’s January formal opening for H200 sales to China, paired with a cap tying China-bound supply to US sales. Nvidia had already said demand was strong but that both governments’ approvals were needed, making Beijing’s import decision the remaining gatekeeper.
The case shows that US export permission is necessary but not sufficient: the transaction can proceed only if Chinese authorities also admit the chips. Later coverage that shipments remained limited to a trivial number of H200s underscores how narrow the channel has been.
First-order effects
- Nvidia can offer a small, licensed volume of H200 chips to identified Chinese customers, but no delivery is assured until Chinese import approval is obtained.
- Chinese prospective buyers gain a potential legal supply route for H200 capacity; the size of that route is constrained by the US license and any Chinese response.
Second-order effects
- Nvidia’s China revenue opportunity remains contingent on two regulatory systems, complicating customer procurement and supply planning even after a US approval.
- The case-by-case licensing framework gives US authorities continued control over shipment volume and customer eligibility, rather than creating an open China market for these chips.
Third-order effects
- If dual approval becomes the norm, advanced-chip trade will be governed less by supplier demand than by interoperating national controls, with market access determined transaction by transaction.
- The split between US export licenses and Chinese import acceptance could make access to AI infrastructure a bilateral policy instrument, not simply a commercial export decision.
The trend: AI-chip access is increasingly being shaped by two-sided geopolitical controls, where an exporter’s license does not by itself create a viable market.